/latentRL

Extracting the latent space of chaotic systems to accelerate convergence of RL controllers.

Primary LanguagePython

latentRL

Next steps

[x] Set up KS solver code

[x] Set up autoencoder training code

[ ] Set up RL code

[x] Decide on some preliminary parameter choices (L, number of actuators etc)

  • $A=5, \sigma=0.1, dt=0.05, T=50, N =256$

[ ] Generate a dataset, for only one L first, then for more

[ ] Train autoencoder

[ ] Train RL agent with and w/o autoencoder